SYNCode (Data)

SYNCode is a collaborative annotation framework that combines human expertise with large language models (LLMs) to improve the quality of annotations for complex, code-centric datasets like Stack Overflow. The system integrates TF-IDF filtering, advanced transformer models (NLP Transformer and UniXcoder), and iterative human refinement to enhance annotation accuracy and reduce bias. A prototype interface enables real-time human-LLM collaboration, demonstrating strong potential for improving annotation reliability in software engineering and NLP applications.